Comparative genomics of Wolbachia and the bacterial species concept

Abstract

The importance of host-specialization to speciation processes in obligate host-associated bacteria is well known, as is also the ability of recombination to generate cohesion in bacterial populations. However, whether divergent strains of highly recombining intracellular bacteria, such as Wolbachia, can maintain their genetic distinctness when infecting the same host is not known. We first developed a protocol for the genome sequencing of uncultivable endosymbionts. Using this method, we have sequenced the complete genomes of the Wolbachia strains wHa and wNo, which occur as natural double infections in Drosophila simulans populations on the Seychelles and in New Caledonia. Taxonomically, wHa belong to supergroup A and wNo to supergroup B. A comparative genomics study including additional strains supported the supergroup classification scheme and revealed 24 and 33 group-specific genes, putatively involved in host-adaptation processes. Recombination frequencies were high for strains of the same supergroup despite different host-preference patterns, leading to genomic cohesion. The inferred recombination fragments for strains of different supergroups were of short sizes, and the genomes of the co-infecting Wolbachia strains wHa and wNo were not more similar to each other and did not share more genes than other A- and B-group strains that infect different hosts. We conclude that Wolbachia strains of supergroup A and B represent genetically distinct clades, and that strains of different supergroups can co-exist in the same arthropod host without converging into the same species. This suggests that the supergroups are irreversibly separated and that barriers other than host-specialization are able to maintain distinct clades in recombining endosymbiont populations. Acquiring a good knowledge of the barriers to genetic exchange in Wolbachia will advance our understanding of how endosymbiont communities are constructed from vertically and horizontally transmitted genes.

Figure 1. Phylogenetic relationships of six Wolbachia strains.

The phylogenetic tree was inferred using the maximum likelihood method from a concatenated alignment of 660 single copy orthologous genes. Numbers on the branches represent the support from 1000 bootstrap replicates. The blue box indicates the strains from supergroup A and the green box indicates the strains from supergroup B.

Figure 2. Genomic overview of the similarity between completely sequenced Wolbachia strains.

Red boxes show the location of supergroup A specific genes (only in the supergroup A strains, wMel, wRi and wHa), whereas green boxes indicate the position of supergroup B specific genes (only in the supergroup B strains wNo and wPip). The yellow boxes represent the location of prophage elements in supergroup A and B strains. The supergroup D strain wBm is included for comparison. Similarity between sequences is indicated by the intensity of the grey lines, where darker is more similar. The lines between the wNo and wHa genomes with the small letters “a” and “b” indicate two regions with atypical synteny and sequence conservation.

Figure 3. Schematic figure of individual tree topologies inferred from single-copy orthologs.

Phylogenetic trees were inferred using the maximum likelihood methods from 660 single-copy gene orthologs, 652 of which supported the division between super-group A and B. The schematic tree summarizes the distribution of the 652 gene trees, where the number of trees supporting a topology within each super-group is shown. The coloured blocks on the right show the number of trees found for each of the 9 possible topologies with separation of the supergroups. The numbers were obtained by clustering individual trees based on weighted Robinson-Fould distances. Numbers in parenthesis represent the trees where all nodes have a bootstrap value higher than 75.

Figure 4. Recombination of chromosomal genes between the supergroup A and B strains.

The yellow block indicates the six genes that produce a gene topology indicative of recombination between wNo and the A-group strains (tree topology shown below the gene order comparison plot). The upstream and downstream flanking genes show the normal separation of the strains into two super-groups (tree topologies shown above the gene order comparison plot).

Figure 5. Horizontal transfer of bacteriophage genes between the supergroup A and B strains.

Phylogenetic trees were inferred using the maximum likelihood methods from four prophage-associated genes A) “Late control D”, B) “Terminase”, C) “Baseplate W”, D) “ANK.”.

Figure 6. Ternary plot of sequence divergence levels at synonymous sites.

The plot shows the variation in synonymous substitution frequencies for the 660 single-copy orthologs in A) supergroup A Wolbachia strains, B) supergroup B Wolbachia strains and C) D. simulans infecting Wolbachia strains. Each dot in the plot represents one gene. Absolute dS-values have been transformed to relative values between 0 and 1 and the mean relative dS-value for each pair is shown on each axis. The spread represents the median distance to the mean point. The color of each dot represents the maximum absolute dS-value among the 3 pairs, ranging from light yellow (low values) to red (high values). The histograms show the frequency of relative dS-values within each pair.

Figure 7. Phylogenetic analysis of the ArgR repressor gene.

The phylogenetic tree was inferred using the maximum likelihood method based on a protein alignment of the ArgR repressor from many different bacterial species. Numbers on the branches represent the support from 1000 bootstrap replicates. The blue box indicates the Wolbachia strains from the supergroup A strains, and the green box indicates Legionella species.